The Roles of Vision and Proprioception
in the Planning of Reaching Movements
Fabrice R. Sarlegna and Robert L. Sainburg
Abstract While vision and proprioception can both provide information about
arm configuration prior to movement, substantial evidence suggests that each
modality is used for different stages of the planning process. In this chapter, we
provide support for the idea that vision is mainly used to define the trajectory
and the kinematics of reaching movements. Proprioception appears to be
critical in the transformation of this plan into the motor commands sent to
the arm muscles.
The idea that information about hand position is critical for accurately
reaching to targets has been supported by many studies over the past few
decades. However, which sensory modalities provide this information, and
how information from these modalities is combined to provide a single estimate
of hand position remains largely unknown. The goal of this chapter is to give an
overview of the research that has addressed how vision and proprioception
might be combined to plan goal-directed arm movements.
It has been well established that a motor plan is prepared in advance of
movement initiation. However, exactly what features of the movement are
explicitly planned remains controversial. It has been suggested that only the
final position of the movement needs to be planned in advance (Feldman and
Latash 2005; Feldman et al. 1998; Polit and Bizzi 1979). However, other studies
have provided evidence that information about the position of the hand prior to
movement is critical to movement accuracy (for a review, Desmurget et al.
1998). For example, Gielen et al. (1985) showed that the peak in movement
acceleration, which occurs so early that it must be specified prior to movement,
scales with intended movement distance. This finding suggests that the central
nervous system estimates movement distance between the starting hand position and the target position to plan the movement. In support of this idea, Ghez
R.L. Sainburg (*)
Department of Kinesiology, Pennsylvania State University, 29, Recreation Building,
University Park, PA, 16802, USA
e-mail: [email protected]
D. Sternad (ed.), Progress in Motor Control,
DOI 10.1007/978-0-387-77064-2_16, Springer ScienceþBusiness Media, LLC 2009
317
318
F.R. Sarlegna and R.L. Sainburg
and colleagues (Ghez and Vicario 1978; Gordon and Ghez 1987a), Gottlieb and
colleagues (Gottlieb et al. 1990, 1996) and Brown and Cooke (1981, 1984, 1986)
showed that various measures of amplitude specification, that arise in the first
50 milliseconds of movement, scale with target amplitude. These include force
amplitude in isometric tasks, and agonist electromyographic amplitude, acceleration amplitude, and torque slope during motor tasks. In addition, Gordon
and Ghez (1987b) provided evidence that errors in specification of these variables can be compensated during the course of movement by extending initial
agonist burst duration, which extends the duration of initial acceleration, or the
time to peak force or torque. This process reflects the use of on-line sensory
information in correcting errors in initial planning. Even though the online
control of movement is deferred by sensory motor loop delays, it is essential if
an unexpected perturbation occurs during the movement (Cordo et al. 1995;
Lestienne 1979; Sainburg et al. 1999; Sarlegna et al. 2004, 2006; Shapiro et al.
2004). For example, if you are going to shake the hand of a friend but he/she
unexpectedly moves his/her hand in a different position, you will, nonetheless,
reach the displaced target, as demonstrated by double-step experiments (e.g.,
Flash and Henis 1991; Goodale et al. 1986; Prablanc and Martin 1992; Sarlegna
2006; Soechting and Lacquaniti 1983). Nevertheless, in this chapter, we will
focus on the planning stage of goal-directed arm movements.
Proprioception Plays a Critical Role in the Planning
of Goal-Directed Movements
Proprioception refers to the component of somatosensation that provides
information about the orientation and motion of body segments and the state
of the muscles. This information is conveyed largely through muscle spindles
and Golgi tendon organs. To investigate the role that proprioception plays
in planning and controlling movements, proprioception has been removed
through either surgery or by studying rare sensory neuropathy conditions, or
has been distorted by the use of vibration. Following the pioneering findings of
Mott and Sherrington (1895), many studies have examined the role of proprioception in movement control by examining movements in non-human primates
who have undergone partial dorsal rhizotomy surgery, including the studies of
Gilman et al. (1976), Taub et al. (1975), Polit and Bizzi (1979) and Gauthier and
Mussa Ivaldi (1988), among others. It has been well established that animals
proprioceptively deafferented in this way show substantial impairments in
accuracy and coordination during multijoint reaching and natural unrestricted
movements. Nevertheless, Polit and Bizzi (1979) demonstrated that performance of single joint targeted elbow movements remained relatively unimpaired, following deafferentation. This brought into question whether
proprioception is necessary to plan and execute movements. In this study
(Polit and Bizzi 1979), monkeys were well trained in the task prior to surgery,
The Roles of Vision and Proprioception in the Planning of Reaching Movements
319
suggesting that they might have been able to recall, but not necessarily plan the
learned movement patterns following deafferentation. Thus, proprioception
may have been crucial for the planning and learning of new movements, but
not for recall of learned patterns. In fact, Polit and Bizzi (1979) also reported
that when the shoulder configuration was altered by rotating the trunk prior to
movements, the movements became grossly inaccurate and uncoordinated.
This supports the idea that the animals recalled a learned plan that was only
appropriate to the perceptual and mechanical conditions experienced during
learning. The monkeys were apparently unable to adapt that plan to changes in
limb configuration. These findings support the idea that proprioception is
essential for providing information about the initial state of the motor
apparatus.
Deafferentation in humans has been studied in the rare condition of largefiber sensory neuropathy, in which damage occurs only to the large diameter
sensory neurons that carry proprioceptive information. Quantitative analyses
of the motor behavior of these patients have emphasized the importance of
proprioception in performing coordinated, accurate movements (Cooke et al.
1985; Forget and Lamarre 1987; Ghez et al. 1995; Gordon et al. 1995; Messier
et al. 2003; Rothwell et al. 1982; Sanes et al. 1985; Sainburg et al. 1993, 1995;
Sarlegna et al. 2006). While deafferented patients can perform many activities
of daily living, given intense concentration and effort (Cole 1995), the quality
and efficiency of performance is severely impaired and patients are often substantially disabled because of this sensory deficit.
Similar to the effects of surgical deafferentation in monkeys, these patients
are able to perform constrained single joint movements without substantial
errors, even in the absence of vision (e.g., Nougier et al. 1996; Sainburg et al.
1995). However, they show severe limitations in the performance of unconstrained activities of daily living, suggesting that proprioception might be
critical for the intersegmental coordination requirements of more complex
movements. Sainburg, Poizner and Ghez (1993) tested this hypothesis by
analyzing the performance of the unconstrained pantomimed gesture of slicing
a loaf of bread, which required sequential out-and-back motion of the hand
within a single plane. Kinematic analysis revealed that even with eyes closed,
healthy subjects performed straight, planar hand movements in the inward and
outward phases, such that the hand paths nicely overlapped. In contrast, movements of deafferented patients were spatially disordered, and severely nonplanar. Analysis of segment kinematics showed that, at the movement reversals,
the elbow and shoulder joint motions were temporally decoupled, causing a
systematic deviation of the movement trajectory. Because intersegmental interaction torques that occur between the moving segments of the limb are maximized at movement reversals (Hoy and Zernicke 1986; Schneider et al. 1989), it
seemed plausible that patients without proprioception were unable to control
these interactions. This might explain the relative accuracy with which patients
could perform single joint motions that do not involve these interactions.
320
F.R. Sarlegna and R.L. Sainburg
Fig. 1 Representative hand paths from control subjects and deafferented patients, drawn
overlaying the template lines, shown for five directions
In order to test this idea, Sainburg et al. (1995) designed a multijoint reaching task that required variations in intersegmental interaction torques. The
task consisted of out-and-back movements made in the horizontal plane
between a fixed starting position and peripheral visual targets. As shown in
Fig. 1, control subjects had no difficulty making straight and overlapping outand-back movements. However, the deafferented patients movements were
substantially more curved and most significantly, showed large errors at movement reversals that depended on the direction of movement. Inverse dynamics
analysis revealed that these errors resulted from the direction dependent variation of interaction torques, which was accounted for by control subjects but
not patients. Sainburg et al. (1995) thus suggested that proprioceptive information during movement was used to update an internal model of limb mechanics,
used to anticipate impending mechanical interactions during the course of
movement.
Proprioception appears critical for controlling other aspects of limb
mechanics, in addition to the effects of intersegmental dynamics. For example,
Ghez et al. (1995) showed that deafferented patients were unable to account for
direction dependent variations in limb inertia during reaching movements,
which resulted in substantial errors in movement distance and direction.
Because direction dependent variations in limb inertia vary with limb configuration, this finding underscored the importance of proprioception in providing configuration information.
The critical role of proprioception in providing initial information of the
postural state of the motor apparatus for movement planning is also supported
by a study of Larish et al. (1984) in healthy subjects. This study showed that
when the limb was vibrated in the absence of visual feedback prior to the
movement, final position was systematically altered. The muscle vibration
presumably biased subjects’ perception of the arm position as the arm would
be perceived as being displaced (DiZio et al. 1993; Lackner and Taublieb 1984).
Larish et al. (1984) therefore suggested that terminal errors increased because
the initial limb configuration was not correctly determined.
The Roles of Vision and Proprioception in the Planning of Reaching Movements
321
Vision is Used to Enhance the Planning of Reaching Movements
Whereas proprioception appears critical for controlling many aspects of
targeted movements, vision can also provide information about the body
segments, as well as essential information about target location. In order to
assess the role of vision to the control of movements, early studies simply
removed visual feedback. One century ago, Woodworth (1899) showed that
movements toward visual targets are more accurate with vision than without
vision. However, whether information about the starting limb position, the
moving limb, or the target was most critical in assuring movement accuracy
was not assessed. More recently, the roles of these different parameters have
been more clearly differentiated (Bagesteiro et al. 2006; Brenner and Smeets
2003; Paillard 1996; Prablanc et al. 1979a; Proteau et al. 2000; Rossetti et al.
1995; Sarlegna et al. 2003, 2006). Vision of the starting hand position has been
specifically manipulated to determine its role in the planning of arm movements. Several studies showed that reaching movements made in the absence of
visual feedback of the moving limb are more accurate when vision of the
starting hand position is available, compared to when such information is
removed (Desmurget et al. 1997; Elliott et al. 1991; Ghez et al. 1995; Prablanc
et al. 1979b; Vindras et al. 1998; for a review, Desmurget et al. 1998). In fact,
when visual feedback of initial position is distorted, systematic errors in movement endpoint are produced (Bagesteiro et al. 2006; Holmes and Spence 2005;
Rossetti et al. 1995; Sainburg et al. 2003; Sarlegna and Sainburg, 2006; Sober
and Sabes 2003). However, these studies have underscored the role of proprioception in contributing to initial position information, by showing that subjects
do not rely completely on visual information about hand position to plan
movements. These findings introduce the question of how vision and propriocpetion might be integrated to provide a single perception of limb position.
Recent findings from our laboratory have suggested that each modality
might provide a distinct and independent estimate of limb position, each
employed for different features of movement planning. Some studies have
suggested that visual information is critical in determining static limb position
because, when vision is removed, judgements about limb position have been
shown to ‘‘drift’’ over time (Paillard and Brouchon 1968; Wann and Ibrahim
1992). Brown, Rosenbaum and Sainburg (2003a, b) directly investigated
whether the tendency of actual limb position to drift during repetitive movements results from a drift in proprioceptive position sense. A schematic of the
experimental set-up is shown in Fig. 2. Subjects are comfortably seated and are
asked to point to visual targets. These targets are presented on a horizontal
screen that prevents direct view of the limb. The recorded coordinates of the
index fingertip are used to project on this screen a cursor that reflects the
position of the hand.
In the experiment of Brown et al. (2003a), subjects were asked to perform a
series of continuous out-and-back movements of the hand between two visual
322
F.R. Sarlegna and R.L. Sainburg
Fig. 2 Experimental set-up
targets, always visible on the screen. The repetitive movements were paced with
a regular audible tone, so that movement speed was similar throughout the
experiment. Subjects had continuous visual feedback of hand position for 5 outand-back cycles, and on the 6th cycle, visual feedback was turned off until the
end of the two minute trial. Subjects were instructed to continue on-pace as
accurately as possible.
Over the 75 movements made during a two-minute repetitive forward-and
backward-cyclic movement trial, the start location of each forward movement
was quantified in order to measure potential drift in limb configuration. This
position drifted an average of 8 centimeters from the initial start location, as
shown in Fig. 3A. This drift increased sharply when visual feedback was removed,
then reached a plateau after approximately 40 trials. The drift was associated with
substantial changes in limb configuration, i.e., elbow and shoulder joint angles
varied substantially. During this drift, the distance and direction of each cycle of
movement remained fairly constant (Fig. 3C). Inverse dynamics analysis, which
yields joint torque estimates from kinematic data, revealed that the torques
required to produce each cycle of movement changed significantly throughout
the session. These changes implied that the neural commands to the muscles took
account of the changes in limb position associated with the drift, in order to
maintain constant movement direction and distance. In fact, forward dynamic
simulations (Fig. 3B) showed that if subjects had not altered their torque strategies to accommodate changes in limb configuration during the course of the drift,
movement directions would have systematically shifted as much as 308 over the
course of the trial.
These findings point out an interesting paradox: Subjects drifted their limb
positions during the course of motion, but apparently accounted for the associated changes in limb configuration in order to maintain movement direction
and distance. This surprising result was confirmed in several conditions, as the
The Roles of Vision and Proprioception in the Planning of Reaching Movements
323
Fig. 3 Drift task from Brown et al. 2003a. A. A single 75 cycle trial. Only outward sections of
movement are shown because analysis was based on this portion of movement. For the first
five cycles, visual feedback of hand position was shown as a screen cursor, and then turned off
prior to cycle 6. Movements progressively drifted to the left, following this. Trial number is
reflected by gray scaling, as indicated by the bar at the left. B. Early movements and late
movements in a single trial did not vary in terms of direction or distance, regardless of drift.
The figure shows the outward portion of an early and a late cycle of an actual trial. The joint
torques calculated from the early movement were used to drive a forward simulation from the
drifted limb position, at the onset of the late movement. The simulated trial indicates the
direction error expected if the altered limb configuration is not accounted for in the joint
torque strategy. Thus, subjects clearly account for the drifted limb configuration in generating
movement strategies. C. Across subjects, movement distance and direction do not change over
subsequent cycles of motion, regardless of drift in limb configuration. Values for each of three
start positions workspace are shown
series of movements were produced from three start locations, in two movement
directions and at three different speeds (Brown et al. 2003a, b). Participants
were able to maintain accurate reproduction of each cycle of the movement
trajectory, even though they failed to maintain their limb position. Thus,
proprioceptive information about limb configuration must have been accurate,
indicating that position drift cannot result from changes in proprioceptive
position sense.
These findings lead to the question of why subjects allowed their limbs to
drift if they had the necessary configuration information to generate joint
324
F.R. Sarlegna and R.L. Sainburg
torques that compensated the associated changes in limb position. The answer
to this question may be that different modalities are employed for regulating
different features of movement. Vision is likely to be the main sensory source of
information to generate a spatial plan for movements directed toward visual
targets. Proprioception, on the other hand, appears to be crucial for transforming this kinematic plan into the variables that correspond to the forces required
to produce the desired kinematics. Thus, both modalities provide information
about initial limb configuration, as required to plan motion, but each modality
appears to inform different features of the control process.
The Relative Contributions of Vision and Proprioception
to the Control of Goal-Directed Arm Movements
The separation of visual and proprioceptive contributions to different features
of control appears to reflect the type of coordinate system in which the information is represented. Vision provides extrinsic, world based coordinate information and is used to plan spatial features of movements toward visual targets.
Proprioception provides intrinsic information about limb configuration and
movement, as well as muscle state, and predominates in transforming a spatial
plan into commands that result in muscle forces and joint torques. Nevertheless, information from both modalities is likely to be combined for each of these
processes, to some extent (Graziano et al. 2000; Ramachandran et al. 1995; van
Beers et al. 1999). It remains unclear exactly how information represented in
disparate coordinates might be combined to yield integrated percepts.
It is interesting to note that Lackner and Taublieb (1984) observed that the
illusory effects of arm muscle vibration (i.e., perceived motion of the arm)
decreased as a function of the amount of visual information (finger only, hand
only or complete vision). This suggests that vision might trump proprioception in
providing configuration information, as proposed initially by Hay et al. (1965).
However, Lackner and Taublieb (1984) emphasized that the vibration-induced
illusion of motion is still observed in full vision conditions, even though it is
attenuated. In fact, DiZio et al. (1993) showed that when a light-emitting diode
(LED) is attached to the finger of subjects in complete darkness, muscle vibration
not only induces illusory motion of the finger but also produces an illusory
motion in vision of the LED. This demonstrates that proprioception can also
trump vision in certain task contexts.
It is plausible that when vision and proprioception are both available,
proprioception provides information about the status of the limb, while vision
provides information about the extrapersonal environment. Prablanc and
Martin (1992) examined the role of the visual feedback about hand position
in making online corrections to target position changes. The visual target was
randomly shifted at movement onset, requiring subjects to modify the direction
of movement. Subjects were able to make rapid adjustments in order to correct
The Roles of Vision and Proprioception in the Planning of Reaching Movements
325
toward the ‘‘new’’ target, highlighting the importance of visual information of
target position. However, the timing or efficacy of these responses was the same,
whether or not subjects had feedback about hand position .(see also Goodale
et al. 1986). However, in a similar target-shift paradigm, Sarlegna et al. (2003)
recently reported that visual feedback of the hand induced a 70 millisecond
advantage over the no-vision condition. In this experiment, the target-shift
required changes in movement distance, rather than direction. This brings
into question whether the role of vision in making on-line movement corrections varies depending on the nature of the required response. Whereas direction corrections in the Prablanc and Martin (1992) study occurred in the initial
acceleration phase of motion, distance corrections in the Sarlegna et al. (2003)
study occurred substantially later, in the deceleration phase. These results
indicate that corrections in distance might evoke longer preparation times
than corrections in direction. It is possible that visual feedback might not yet
be available for the early, direction corrections, due to visual-motor loop
delays. However, recent studies have shown visual-motor latencies of under
200 milliseconds (Day and Lyon 2000; Desmurget et al. 1999; Sarlegna et al.
2004; Saunders and Knill 2004), which renders visual motor loop delay as a
poor candidate to explain the differences in visual contributions between
responses to distance and direction shifts in the visual target. More research is
necessary to distinguish whether visual feedback differentially mediates corrections in movement direction and distance, or whether the observed differences
in visual contributions to movement corrections might result from limitations in
visual-motor loop delays.
It should be stressed that completely removing information from a given
modality may not provide accurate information about the relative contribution
of that modality to the control process. For example, when vision is removed,
our ability to reach for objects accurately should not be interpreted as a
uselessness of visual feedback in controlling movements but rather as an adequacy of proprioception to mediate the coordination of volitional movements.
As Cruse et al. (1990) pointed out, it is remarkable that elimination of visual
input can have minor effects on motor behavior, while distortion of the very
same signal can have drastic effects. In line with this idea, we will present how
visual distortion techniques have helped us to better understand its relative
contribution to the control of purposeful actions. In a seminal study using such
techniques, Hay et al. (1965) dissociated visual and proprioceptive information
by using prismatic goggles, which rotated a subject’s visual field by a few
degrees. This dissociated the ‘‘seen’’ position of the hand from its ‘‘felt’’ position.
Interestingly, subjects indicated that they did not notice any mismatch, and that
they felt their hand where they saw it. More recently, Rossetti, Desmurget and
Prablanc (1995) used prisms to displace the location of an LED, representing
hand position prior to movements. The LED was turned off at movement onset,
and movements were performed in the dark, eliminating visual feedback about
hand motion during the course of movement. The results indicated that the
direction of movement was biased toward the distorted visual information of
326
F.R. Sarlegna and R.L. Sainburg
the starting hand position, and that proprioception also played a significant
role in determining movement direction. In short, subjects reached as though
their hand was located between its seen and actual start location. The authors
concluded that visual and proprioceptive information are fused together to
provide a single estimate of hand location. However, recent research from our
laboratory, as well as others, has suggested that both modalities provide independent information to different components of motor planning, a hypothesis
that we detail in the next section of this chapter.
Visual and Proprioceptive Contributions to Two Serial Stages
of Motor Planning
Visual perception appears to influence the way movements are planned. For
example, Morasso has shown in several studies that during goal-directed arm
movements, the hand tends to move fairly straight, even though the joint
displacement patterns associated with such movements can be quite complex
and even change direction (e.g., Morasso 1981). In fact, Flanagan and Rao
(1995) and Wolpert et al. (1995) showed that when visual feedback of apparent
hand position was displayed in a non-linear joint angle-based coordinate system, subjects arm movements became very curved in order to make straight
cursor motions. Together, these findings support the idea that hand movements
are planned as vectors, defined in extrinsic, world based coordinates. In fact,
several studies have indicated that movement distance and direction, defined in
extrinsic coordinates, are separately planned prior to targeted movement (Ghez
et al. 1997; Rosenbaum 1980). Supporting the idea that movements are initially
planned in extrinsic, visual space, studies of reaching in non-human primates
have shown that motor cortex neuron firing patterns correlate well with hand
direction, defined in extrinsic coordinates (Georgopoulos et al. 1986; for a
review, Georgopoulos 1998; see also Alexander and Crutcher 1990; Fu et al.
1995; Kakei et al. 1999; Sergio and Kalaska 2003). However it should be
emphasized that other studies have shown that intrinsic variables, such as
joint torque, might show as good or better correlations with motor cortex
neuron activities (Scott and Kalaska 1995).
While goal directed arm movements toward visual targets appear to be
planned according to an extrinsic coordinate system, as sampled by vision,
ultimately this plan must be transformed into variables that correspond to the
forces required to produce motion. Thus, a plan to move the hand toward a
target must be transformed into the motor commands that ultimately result in
muscle activations, a transformation akin to the computational process of
inverse dynamics. Interestingly, Atkeson and Hollerbach (1985) showed that
unconstrained movements, performed at different target speeds and with
different loads, exhibit similar shaped, unimodal hand tangential velocity profiles (i.e., the same kinematics). The joint coordination patterns associated with
The Roles of Vision and Proprioception in the Planning of Reaching Movements
327
these simple hand velocities were often quite complex and varied across movements. This suggests that the movement kinetics were modified so that a reliable
kinematic plan could be executed under varied dynamic conditions. Flash and
Hogan (1985) presented a model for the organization of arm movements that
attempted to explain the aforementioned characteristic features of kinematic
profiles. They suggested that subjects might be constraining the kinematic plan
to the smoothest possible movement. Smooth movements are characterized by
low jerk, the 4th derivative of displacement. When Flash and Hogan performed
optimizations for a large variety of two segment arm movements, the solutions
were straight hand paths and unimodal, bell shaped velocity profiles, similar to
the profiles characterized in human subjects (Papaxanthis et al. 2005). Thus,
limb segment kinetics appear to be modified in order to make smooth and
efficient movements of the hand as it is transported to the target.
These studies suggested that movement planning might take place in two
major phases: First, in the initial stage, initial limb configuration, target position and the environmental conditions would be taken into account to determine a kinematic plan for the desired movement (e.g., at a very high speed, with
a curved trajectory to avoid an obstacle). Then, a second stage of planning
might correspond to the transformation of that plan into the motor commands
that result in the forces producing motion. Supporting evidence for this view
comes from a study (Sainburg et al. 2003) in which subjects pointed to visual
targets using a similar set-up as the one shown in Fig. 1. A virtual-reality system
was used to introduce a mismatch between the actual fingertip position and its
seen position (the cursor representing hand position) prior to reaching movements. In this manner, it was possible to dissociate visual from proprioceptive
information about starting position. Cursor start position and finger start
position differed by as much as 5 cm across nine different fingertip positions
that corresponded to a single cursor start position (see Fig. 4). The analysis of
movement trajectories showed that movements from all fingertip start locations
were matched in terms of direction to the vector defined by the cursor start
location and the target. Movement direction was thus planned on the basis of
visual information of starting position while proprioceptive information did
not appear to contribute to the directional planning. Inverse dynamics analysis
revealed that muscle torques changed substantially between the different starting arm configurations. Thus, subjects adjusted their muscle torques so that
movement direction matched the direction of the vector between the cursor
start location and the target. This compensation suggested that proprioception
provided accurate information about arm configuration, supporting the findings of Brown et al. (2003a). We thus concluded that the direction of movement
was planned in accord with visual start location, while the dynamic strategy
employed to execute this plan was based on proprioceptive information.
The fact that movement direction appeared to be planned in visual coordinates (on the basis of vision of the cursor and the target) was confirmed in a
subsequent study from our laboratory (Lateiner and Sainburg 2003). In this
study, a single hand position was used but the cursor could be initially presented
328
F.R. Sarlegna and R.L. Sainburg
Fig. 4 Ensemble averaged hand paths – standard errors in X and Y coordinates (crosses) are
shown for all nine start positions of the hand. The initial position of the cursor was always on
the central start, while the starting position of the finger varied between the central start
location (veridical, baseline trials) and each of 8 other positions, as shown
either at a location corresponding to the veridical hand position or at a location
differing by a few centimeters. Using such a visuo-proprioceptive mismatch
confirmed that visual information of hand and target positions was used to plan
movement direction. Sober and Sabes (2003) used a similar experimental protocol to examine the relative contributions of vision and proprioception to the
planning of movement direction toward visual targets. However, they assessed
the findings using a computational model in which each modality could contribute to each of two different stages of the planning process: Direction planning in extrinsic coordinates, and the transformation of this plan into intrinsic,
dynamic related variables. The model included coefficients that provided a
weighted contribution of each modality. According to this model, the first
stage of direction planning relied predominantly on visual information
(80%), whereas the second stage of processing, the transformation of the
direction plan into joint-based coordinates, relied mostly on proprioceptive
information (70%). This result was consistent with those of Sainburg et al.
(2003) and Lateiner and Sainburg (2003) indicating that the extrinsic planning
stage of movement direction appears to rely mainly on visual information of
starting hand location.
Whereas the studies reviewed above investigated the processes underlying
direction planning, the planning of movement distances might employ processes quite distinct from those employed in planning movement directions
(Fu et al. 1995; Ghez et al. 1991, 1997; Paillard 1996; Riehle and Requin
1989; Rosenbaum 1980; Soechting and Flanders 1989). In a series of studies,
(Bagesteiro, Sarlegna and Sainburg 2006; Sarlegna and Sainburg 2007), we
investigated the relative roles of proprioception and vision to the control of
movement distance, in a single joint task that has no requirements for directional planning. A visuo-proprioceptive mismatch of initial hand position was
introduced prior to the reaching movements, performed without visual
The Roles of Vision and Proprioception in the Planning of Reaching Movements
329
feedback of the limb once the movement started. Figure 5 shows sample hand
paths (5A) under three randomly presented conditions: The left column shows
movements performed under veridical conditions, in which the initial positions
of the cursor and hand were the same. In all cases the target (large open circle)
was the same. However, either the cursor (small closed circle) or the hand
position (beginning of hand path) was different. The conditions were presented
randomly, and subjects reported that they were not aware of any dissociations
or odd conditions in during the course of the study. As can be readily determined from the figure, when the initial positions of cursor and finger were the
same (Fig. 5A, left column), movements were adjusted, such that start position
3 elicited longer movements than start position 1. This finding was consistent
Fig. 5 Representative hand paths A. from single joint movements toward a single target under
three start locations when the position of the cursor and the finger were the same (left column),
when the cursor position varied, but the finger position was constant across start positions
(center column), and when the cursor position was constant across start positions, while the
finger position varied (right column). Bar graphs B. show the corresponding distances,
averaged across subjects, for the conditions shown in A
330
F.R. Sarlegna and R.L. Sainburg
across subjects, as reflected by the bar graph in figure 5B, left column. When the
finger started from the same location, but the cursor was shown in the varied
start positions (Fig. 5A, center column), subjects also varied movement distance, such that the finger substantially overshot the target, when the starting
cursor location was shown in position 3, farthest from the target. In other
words, subjects appeared to plan target distance, based predominantly on the
seen location of the cursor, rather than on the felt location of the hand. Nevertheless, the location of the hand did contribute, though less substantially, to
distance planning. This is reflected by the hand paths in the right column of
Fig. 5A. While the initial cursor location was the same between the three
conditions, the finger location varied. The distance of the movements did
vary slightly across conditions, as though the cursor position, and not the
finger position, was the major determinant of movement distance. These
findings are quantified across subjects in Fig. 5B, which shows the greatest
change in distance under veridical conditions (left: 9.9 centimeters), followed
by the condition in which the cursor but not the hand changed its initial
location (center: 5.5 cm). Finally, the least variation in distance occurred for
the conditions in which the initial cursor location was constant, while the
initial finger location varied (right: 3.4 cm). Interestingly, the summed variation in movement distance produced by varying cursor location and hand
location (8.9 cm) was very similar to that produced by varying both hand and
cursor location, veridically (9.2 cm), as though the two modalities contributed
independently to distance planning.
These findings indicated that movement distance was planned predominantly on the basis of the visually displayed distance between the cursor and
the target (Fig. 5). This was emphasized by the fact that peak acceleration was
scaled with initial cursor position, but not with initial hand location (Sarlegna
and Sainburg 2007). These findings indicated that, for each starting cursor
position, the initial movement acceleration (and therefore the initial joint
torque amplitude) remained constant, regardless of variations in elbow starting angle that were as large as 158. Because of configuration dependent variations in muscle moment arm and length-tension relations, production of the
same torque across different joint angles requires substantial changes in the
neural commands to the muscles. In fact, Prodoehl et al. (2003) recently
showed systematic and predictable changes in joint torque during maximum
speed tasks across the range of elbow movements that we employed in our
study. In Prodoehl et al.’s study, electromyographic (EMG) activity of elbow
muscles showed compensatory changes in timing and amplitude that served to
diminish the effect of configuration dependent changes in muscle moment arm
and muscle length. In a similar study of ankle joint movements, Nourbakhsh
and Kukulka (2004) showed that when subjects produced targeted isometric
plantar flexion torques across a range of ankle positions, triceps surae EMG
was modulated, so as to compensate configuration dependent variations in
muscle moment arm.
The Roles of Vision and Proprioception in the Planning of Reaching Movements
331
We thus expect that maintaining constant acceleration amplitudes over
experimentally imposed variations in starting elbow angle must have induced
compensatory variations in muscle activation strategies. These findings indicate
that while subjects used visual information to plan movement distance toward
visual targets, proprioceptive information about hand position was used to
transform that plan into variables that correspond to the joint torque required
to produce the planned movement. These finding are consistent with previous
research in proprioceptively deafferented patients, which indicated a major role
of proprioception in transforming kinematic plans into dynamic strategies
required for intersegmental coordination (Sainburg et al. 1993, 1995).
Conclusion
In this chapter, we presented experimental evidence for the idea that vision and
proprioception contribute differentially to the planning of goal-directed arm
movements toward visual targets. Visual information of hand position is mainly
used in a first stage to define the kinematic plan of the reaching movement
within an extrinsic coordinate system. Proprioception appears to be essential in
transforming this kinematic profile into the neural commands that result in the
necessary forces to achieve the desired motion.
References
Alexander GE, Crutcher MD (1990) Neural representations of the target (goal) of visually
guided arm movements in three motor areas of the monkey. J Neurophysiol 64: 164–178
Atkeson CG, Hollerbach JM (1985) Kinematic features of unrestrained vertical arm movements. J Neurosci 5: 2318–2330
*Bagesteiro LB, Sarlegna FR, Sainburg RL (2006) Differential influence of vision and
proprioception on control of movement distance. Exp Brain Res 171: 358–370
Brenner E, Smeets JB (2003) Fast corrections of movements with a computer mouse. Spat Vis
16: 365–376
*Brown LE, Rosenbaum DA, Sainburg RL (2003a) Limb position drift: implications for
control of posture and movement. J Neurophysiol 90: 3105–3118
Brown LE, Rosenbaum DA, Sainburg RL (2003b) Movement speed effects on limb position
drift. Exp Brain Res 153: 266–274
Brown SH, Cooke JD (1981) Responses to force perturbations preceding voluntary human
arm movements. Brain Res 220: 350–355
Brown SH, Cooke JD (1984) Initial agonist burst duration depends on movement amplitude.
Exp Brain Res 55: 523–527
Brown SH, Cooke JD (1986) Initial agonist burst is modified by perturbations preceding
movement. Brain Res 377: 311–322
Cole JO (1995) Pride and a Daily Marathon. Paperback, Bradford Books
Cooke JD, Brown S, Forget R, Lamarre Y (1985) Initial agonist burst duration changes with
movement amplitude in a deafferented patient. Exp Brain Res 60: 184–187
332
F.R. Sarlegna and R.L. Sainburg
Cordo P, Bevan L, Gurfinkel V, Carlton L, Carlton M, Kerr G (1995) Proprioceptive
coordination of discrete movement sequences: mechanism and generality. Can J Physiol
Pharmacol 73: 305–315
Cruse H, Dean J, Heuer H, Scmidt RA (1990) Utilization of sensory information for motor
control. In: Neumann O, Prinz W (eds) Relationships between perception and action,
pp 43–79. Berlin: Springer-Verlag
Day BL, Lyon IN (2000) Voluntary modification of automatic arm movements evoked by
motion of a visual target. Exp Brain Res 130: 159–168
*Desmurget M, Epstein CM, Turner RS, Prablanc C, Alexander GE, Grafton ST (1999) Role
of the posterior parietal cortex in updating reaching movements to a visual target. Nat
Neurosci 2: 563–567
Desmurget M, Pelisson D, Rossetti Y, Prablanc C (1998) From eye to hand: planning goaldirected movements. Neurosci Biobehav Rev 22: 761–788
Desmurget M, Rossetti Y, Jordan M, Meckler C, Prablanc C (1997) Viewing the hand prior to
movement improves accuracy of pointing performed toward the unseen contralateral
hand. Exp Brain Res 115: 180–186
DiZio P, Lathan CE, Lackner JR (1993) The role of brachial muscle spindle signals in
assignment of visual direction. J Neurophysiol 70: 1578–1584
Elliott D, Carson RG, Goodman D, Chua R (1991) Discrete vs. continuous visual control of
manual aiming. Hum Mov Sci 10: 393–418
Feldman AG, Latash ML (2005) Testing hypotheses and the advancement of science: recent
attempts to falsify the equilibrium point hypothesis. Exp Brain Res 161: 91–103
Feldman AG, Ostry DJ, Levin MF, Gribble PL, Mitnitski AB (1998) Recent tests of the
equilibrium-point hypothesis (lambda model). Motor Control 2: 189–205
Flanagan JR, Rao AK (1995) Trajectory adaptation to a nonlinear visuomotor transformation:
evidence of motion planning in visually perceived space. J Neurophysiol 74: 2174–2178
Flash T, Henis E (1991) Arm trajectory modification during reaching towards visual targets.
J Cognitive Neurosci 3: 220–230
Flash T, Hogan N (1985) The coordination of arm movements: an experimentally confirmed
mathematical model. J Neurosci 5: 1688–1703
Forget R, Lamarre Y (1987) Rapid elbow flexion in the absence of proprioceptive and
cutaneous feedback. Hum Neurobiol 6: 27–37
Fu QG, Flament D, Coltz JD, Ebner TJ (1995) Temporal encoding of movement kinematics
in the discharge of primate primary motor and premotor neurons. J Neurophysiol 73:
836–854
Gauthier GM, Mussa Ivaldi F (1988) Oculo-manual tracking of visual targets in monkey: role
of the arm afferent information in the control of arm and eye movements. Exp Brain Res
73: 138–154
Georgopoulos AP (1998) Online visual control of the arm. Novartis Found Symp 218: 147–164
Georgopoulos AP, Schwartz AB, Kettner RE (1986) Neuronal population coding of movement direction. Science 233: 1416–1419
Ghez C, Favilla M, Ghilardi MF, Gordon J, Bermejo R, Pullman S (1997) Discrete and
continuous planning of hand movements and isometric force trajectories. Exp Brain Res
115: 217–233
Ghez C, Gordon J, Ghilardi MF (1995) Impairments of reaching movements in patients without
proprioception. II. Effects of visual information on accuracy. J Neurophysiol 73: 361–372
*Ghez C, Hening W, Gordon J (1991) Organization of voluntary movement. Curr Opin
Neurobiol 1: 664–671
Ghez C, Vicario D (1978) The control of rapid limb movement in the cat. II. Scaling of
isometric force adjustments. Exp Brain Res 33: 191–202
Gielen CC, van den Oosten K, Pull ter Gunne F (1985) Relation between EMG activation
patterns and kinematic properties of aimed arm movements. J Mot Behav 17: 421–442
Gilman S, Carr D, Hollenberg J (1976) Kinematic effects of deafferentation and cerebellar
ablation. Brain 99: 311–330
The Roles of Vision and Proprioception in the Planning of Reaching Movements
333
Goodale MA, Pelisson D, Prablanc C (1986) Large adjustments in visually guided reaching do
not depend on vision of the hand or perception of target displacement. Nature 320: 748–750
Gordon J, Ghez C (1987a) Trajectory control in targeted force impulses. II. Pulse height
control. Exp Brain Res 67: 241–252
Gordon J, Ghez C (1987b) Trajectory control in targeted force impulses. III. Compensatory
adjustments for initial errors. Exp Brain Res 67: 253–269
Gordon J, Ghilardi MF, Ghez C (1995) Impairments of reaching movements in patients
without proprioception. I. Spatial errors. J Neurophysiol 73: 347–360
Gottlieb GL, Chen CH, Corcos DM (1996) Nonlinear control of movement distance at the
human elbow. Exp Brain Res 112: 289–297
Gottlieb GL, Corcos DM, Agarwal GC, Latash ML (1990) Organizing principles for single
joint movements. III. Speed- insensitive strategy as a default. J Neurophysiol 63: 625–636
Graziano MS, Cooke DF, Taylor CS (2000) Coding the location of the arm by sight. Science
290: 1782–1786
Hay JC, Pick HL Jr, Ikeda K (1965) Visual capture produced by prism spectacles. Psychon Sci
2: 215–216
Holmes NP, Spence C (2005) Visual bias of unseen hand position with a mirror: spatial and
temporal factors. Exp Brain Res 166: 489–497
Hoy MG, Zernicke RF (1986) The role of intersegmental dynamics during rapid limb
oscillations. J Biomech 19: 867–877
Kakei S, Hoffman DS, Strick PL (1999) Muscle and movement representations in the primary
motor cortex. Science 285: 2136–2139
Lackner JR, Taublieb AB (1984) Influence of vision on vibration-induced illusions of limb
movement. Exp Neurol 85: 97–106
Larish DD, Volp CM, Wallace SA (1984) An empirical note on attaining a spatial target after
distorting the initial conditions of movement via muscle vibration. J Mot Behav 16: 76–83
Lateiner JE, Sainburg RL (2003) Differential contributions of vision and proprioception to
movement accuracy. Exp Brain Res 151: 446–454
Lestienne F (1979) Effects of inertial load and velocity on the braking process of voluntary
limb movements. Exp Brain Res 35: 407–418
Messier J, Adamovich S, Berkinblit M, Tunik E, Poizner H (2003) Influence of movement
speed on accuracy and coordination of reaching movements to memorized targets in threedimensional space in a deafferented subject. Exp Brain Res 150: 399–416
Morasso P (1981) Spatial control of arm movements. Exp Brain Res 42: 223–227
Mott FW, Sherrington CS (1895) Experiments upon the influence of sensory nerves upon.
movement and nutrition of the limbs. Proc R Soc B 57:481–488
Nougier V, Bard C, Fleury M, Teasdale N, Cole J, Forget R, Paillard J, Lamarre Y (1996)
Control of single-joint movements in deafferented patients: evidence for amplitude coding
rather than position control. Exp Brain Res 109: 473–482
Nourbakhsh MR, Kukulka CG (2004) Relationship between muscle length and moment arm
on EMG activity of human triceps surae muscle. J Electromyogr Kinesiol 14: 263–273
*Paillard J (1996) Fast and slow feedback loops for the visual correction of spatial errors in a
pointing task: a reappraisal. Can J Physiol Pharmacol 74: 401–417
Paillard J, Brouchon M (1968) Active and passive movements in the calibration of position
sense. In: Freedman SJ (ed) The Neuropsychology of spatially oriented behaviour
(pp 35–56). Illinois: Dorsey Press
Papaxanthis C, Pozzo T, McIntyre J (2005) Kinematic and dynamic processes for the control
of pointing movements in humans revealed by short-term exposure to microgravity.
Neuroscience 135: 371–383
Polit A, Bizzi E (1979) Characteristics of motor programs underlying arm movements in
monkeys. J Neurophysiol 42: 183–194
Prablanc C, Echallier JE, Jeannerod M, Komilis E (1979a) Optimal response of eye and hand
motor systems in pointing at a visual target. II. Static and dynamic visual cues in the
control of hand movement. Biol Cybern 35: 183–187
334
F.R. Sarlegna and R.L. Sainburg
Prablanc C, Echallier JF, Komilis E, Jeannerod M (1979b) Optimal response of eye and hand
motor systems in pointing at a visual target. I. Spatio-temporal characteristics of eye and
hand movements and their relationships when varying the amount of visual information.
Biol Cybern 35: 113–124
Prablanc C, Martin O (1992) Automatic control during hand reaching at undetected twodimensional target displacements. J Neurophysiol 67: 455–469
Prodoehl J, Gottlieb GL, Corcos DM (2003) The neural control of single degree-of-freedom
elbow movements. Effect of starting joint position. Exp Brain Res 153: 7–15
Proteau L, Boivin K, Linossier S, Abahnini K (2000) Exploring the limits of peripheral vision
for the control of movement. J Mot Behav 32: 277–286
Ramachandran VS, Rogers-Ramachandran D, Cobb S (1995) Touching the phantom limb.
Nature 377: 489–490
Riehle A, Requin J (1989) Monkey primary motor and premotor cortex: single-cell activity
related to prior information about direction and extent of an intended movement.
J Neurophysiol 61: 534–549
Rosenbaum DA (1980) Human Movement Initiation: Specification of arm, direction, and
extent. J Exp Psychol 109: 444–474
Rossetti Y, Desmurget M, Prablanc C (1995) Vectorial coding of movement: vision, proprioception, or both? J Neurophysiol 74: 457–463
Rothwell JC, Traub MM, Day BL, Obeso JA, Thomas PK, Marsden CD (1982) Manual
motor performance in a deafferented man. Brain 105: 515–542
*Sainburg RL, Ghez C, Kalakanis D (1999) Intersegmental dynamics are controlled by
sequential anticipatory, error correction, and postural mechanisms. J Neurophysiol 81:
1040–1056
*Sainburg RL, Ghilardi MF, Poizner H, Ghez C (1995) Control of limb dynamics in normal
subjects and patients without proprioception. J Neurophysiol 73: 820–835
Sainburg RL, Lateiner JE, Latash ML, Bagesteiro LB (2003) Effects of altering initial
position on movement direction and extent. J Neurophysiol 89: 401–415
Sainburg RL, Poizner H, Ghez C (1993) Loss of proprioception produces deficits in interjoint
coordination. J Neurophysiol 70: 2136–2147
Sanes JN, Mauritz KH, Dalakas MC, Evarts EV (1985) Motor control in humans with largefiber sensory neuropathy. Hum Neurobiol 4: 101–114
Sarlegna F, Blouin J, Bresciani JP, Bourdin C, Vercher JL, Gauthier GM (2003) Target and
hand position information in the online control of goal-directed arm movements. Exp
Brain Res 151: 524–535
Sarlegna F, Blouin J, Vercher JL, Bresciani JP, Bourdin C, Gauthier GM (2004) Online
control of the direction of rapid reaching movements. Exp Brain Res 157: 468–471
Sarlegna FR (2006) Impairment of online control of reaching movements with aging: a
double-step study. Neurosci Lett 403: 309–314
Sarlegna FR, Gauthier GM, Bourdin C, Vercher JL, Blouin J (2006) Internally driven control
of reaching movements: a study on a proprioceptively deafferented subject. Brain Res Bull
69: 404–415
Sarlegna F, Sainburg RL (2007) The effect of target modality on visual and propriocaptive
contributions to the control of movement distance. Exp Brain Res 176: 267–280
Sarlegna FR, Sainburg RL, The effect of target modality on visual on visual and proprioceptive contributions to control of movement distance. xp Brain Res. 2006 Aug 4
Saunders JA, Knill DC (2004) Visual feedback control of hand movements. J Neurosci 24:
3223–3234
Schneider K, Zernicke RF, Schmidt RA, Hart TJ (1989) Changes in limb dynamics during the
practice of rapid arm movements. J Biomech 22: 805–817
Scott SH, Kalaska JF (1995) Changes in motor cortex activity during reaching movements
with similar hand paths but different arm postures. J Neurophysiol 73: 2563–2567
The Roles of Vision and Proprioception in the Planning of Reaching Movements
335
Sergio LE, Kalaska JF (2003) Systematic changes in motor cortex cell activity with arm
posture during directional isometric force generation. J Neurophysiol 89: 212–228
Shapiro MB, Gottlieb GL, Corcos DM (2004) EMG responses to an unexpected load in fast
movements are delayed with an increase in the expected movement time. J Neurophysiol
91: 2135–2147
Sober SJ, Sabes PN (2003) Multisensory integration during motor planning. J Neurosci 23:
6982–6992
Soechting JF, Flanders M (1989) Sensorimotor representations for pointing to targets in
three- dimensional space. J Neurophysiol 62: 582–594
Soechting JF, Lacquaniti F (1983) Modification of trajectory of a pointing movement in
response to a change in target location. J Neurophysiol 49: 548–564
Taub E, Goldberg IA, Taub P (1975) Deafferentation in monkeys: pointing at a target
without visual feedback. Exp Neurol 46: 178–186
van Beers RJ, Sittig AC, Gon JJ (1999) Integration of proprioceptive and visual positioninformation: An experimentally supported model. J Neurophysiol 81: 1355–1364
Vindras P, Desmurget M, Prablanc C, Viviani P (1998) Pointing errors reflect biases in the
perception of the initial hand position. J Neurophysiol 79: 3290–3294
Wann JP, Ibrahim SF (1992) Does limb proprioception drift? Exp Brain Res 91: 162–166
*Wolpert DM, Ghahramani Z, Jordan MI (1995) Are arm trajectories planned in kinematic
or dynamic coordinates? An adaptation study. Exp Brain Res 103: 460–470
Woodworth RS (1899) The accuracy of voluntary movement. Psychol Rev 3: 1–114
The references marked with an asterisk (*) are specifically recommended for further introduction or background to the topic.